Sinclair Broadcast Group: Video Intelligence Using Machine Learning
Use Machine Learning to Identify Relevant Content for Users
Sinclair is one of the largest and most diversified television broadcasting companies in the country. The Company owns, operates and/or provides services to 191 television stations in 89 markets. Sinclair is a leading local news provider in the country and is dedicated to impactful journalism with a local focus. The Company has multiple national networks, live local sports production, as well as stations affiliated with all the major networks. Sinclair’s content is delivered via multiple platforms, including over-the-air, multi-channel video program distributors, and digital platforms.
Sinclair generates hundreds of gigabytes of video assets a day. They want to adopt ML to intelligently produce content that are relevant for consumers. For example, identifying start and end timings of an hour long news segment. This will allow their consumers to get to the specific content they are looking for quickly. Other use cases they were interested in included adding metadata such as person names, places in order to make the video content searchable.
Google Cloud Implementation
Sinclair partnered with Google Cloud Platform (GCP) and SpringML to build a solution that initially focused on automatically identifying a weather segment in news videos. This solution was developed using GCP’s Machine Learning API’s such as Video intelligence API, Speech API, and Vision API in conjunction with custom Tensorflow models. The resulting model was also integrated into their video streaming system so that such weather highlights can be made available as soon as the news bulletin finishes.
Sinclair was able to generate highlights for the tennis channel videos using Machine learning.